The Spectral Decomposition Estimate Of Panel Data With Mixed Effect Model

The multidimensional data analysis methods of Panel Data are now the frontier of statistical science. Panel Data is a typical mixed linear model with the variables of time and individual. Its research development mainly focuses on the random effect model and fixed effect model independently. The debate between fixed effect and random effect has been haunting in Panel Data analysis since its origin. The inaccurate model brings on the inefficacy of the parameters estimate. So this thesis' purpose is to design a new mixed effect model with individual and time random effects and simultaneously with fixed and error effect together. With this new model structure, I introduce a new parameter estimation method, called as the spectral decomposition estimate.This is a new and more scientific model structure, and this method is also a new method to this model. It is a certain improvement to the traditional Panel Data research methods and has some excellent characteristics.